cypress-allure-plugin

Plugin for integrating allure reporter in Cypress with support of Allure API.

Installation

• Allure binary: directly from allure2 or allure-commandline npm package.

• Java 8 (required to run allure binary)

• There is no need to set this plugin as reporter in Cypress or use any other allure reporters. Just download:

  • using yarn:

  yarn add -D @shelex/cypress-allure-plugin

  • using npm:

  npm i -D @shelex/cypress-allure-plugin
  

Setup

Cypress v10+

• Use defineConfig and setupNodeEvents inside config.js\config.ts files:

  const allureWriter = require('@shelex/cypress-allure-plugin/writer');
  // import allureWriter from "@shelex/cypress-allure-plugin/writer";
  
  module.exports = defineConfig({
      e2e: {
          setupNodeEvents(on, config) {
              allureWriter(on, config);
              return config;
          }
      }
  });

  • if you have webpack or other preprocessors

    • Please take into account that some plugins/preprocessors may register event listeners in Cypress (especially after:spec to have access to results) which will block other plugins - cypress#22428. To make allure-plugin work with such plugins/preprocessors please use env variable allureReuseAfterSpec: true, it will listen to cypress process directly, avoiding registering a listener (but if none of listeners is registered - process will not receive an event).

    • use defineConfig and setupNodeEvents inside config.js\config.ts files:

    const allureWriter = require('@shelex/cypress-allure-plugin/writer');
    // import allureWriter from "@shelex/cypress-allure-plugin/writer";
    
    module.exports = defineConfig({
        e2e: {
            setupNodeEvents(on, config) {
                on('file:preprocessor', webpackPreprocessor);
                allureWriter(on, config);
                return config;
            },
            env: {
                allureReuseAfterSpec: true
            }
        }
    });

• Register commands in cypress/support/e2e.js file:

  • with import:

  import '@shelex/cypress-allure-plugin';

  • with require:

  require('@shelex/cypress-allure-plugin');

Cypress before v10

• Connect plugin in cypress/plugins/index.js. Take into account that Cypress generate plugins file with module.exports on the first initialization but you should have only one export section. In order to add Allure writer task just replace it or add writer task somewhere before returning config:

  const allureWriter = require('@shelex/cypress-allure-plugin/writer');
  // import allureWriter from "@shelex/cypress-allure-plugin/writer";
  
  module.exports = (on, config) => {
      allureWriter(on, config);
      return config;
  };

  • if you have webpack or other preprocessors

    • please set allure writer before returning "config":

    const allureWriter = require('@shelex/cypress-allure-plugin/writer');
    // import allureWriter from "@shelex/cypress-allure-plugin/writer";
    
    module.exports = (on, config) => {
        on('file:preprocessor', webpackPreprocessor);
        allureWriter(on, config);
        return config;
    };

• Register commands in cypress/support/index.js file:

  • with import:

  import '@shelex/cypress-allure-plugin';

  • with require:

  require('@shelex/cypress-allure-plugin');

Autocompletion

• for IntelliSense (autocompletion) support in your IDE add tsconfig.json and specify types property for compilerOptions:

 "compilerOptions": {
        "allowJs": true,
        "baseUrl": "./",
        "types": ["@shelex/cypress-allure-plugin"],
        "noEmit": true
    },

Configuration

Plugin is customizable via Cypress environment variables:

General

Steps

Attachments

Links

No idea

These options could be passed in multiple ways, you can check docs.
But also you can use allure.properties file (however allure=true is still required to be passed as cypress env variable):

allure.results.directory=allure-results
allure.link.issue.pattern=https://example.com/project/test/issue/
allure.link.tms.pattern=https://example.com/testcases/TEST/
allure.reuse.after.spec=false
allure.clear.skipped=false
allure.cypress.log.commands=true
allure.cypress.log.gherkin=true
allure.cypress.log.requests=true
allure.skip.automatic.screenshot=false
allure.omit.previous.attempt.screenshot=false
allure.video.passed=false
allure.analytics=false

Execution

• To enable Allure results writing just pass environment variable allure=true, example:

npx cypress run --env allure=true

• if allure is enabled, you can check data in runtime, just send it to Chrome Developer tools console:

Cypress.Allure.reporter.runtime.writer;

How to open report

Assuming allure is already installed:

• serve report based on current "allure-results" folder: allure serve
• generate new report based on current "allure-results" folder: allure generate
• open generated report from "allure-report" folder: allure open

Examples

See cypress-allure-plugin-example project, which is already configured to use this plugin, hosting report as github page and run by github action. It has configuration with complete history (allure can display 20 build results ) with links to older reports and links to CI builds.

There are also existing solutions that may help you prepare your report infrastructure:

• Allure docker service - highly customizable feature-rich container
• Allure Server - self-hosted portal with your reports
• allure-reports-portal - another portal which allows to gather reports for multiple projects in single ui
• allure-static-booster - solution for generating self-hosted Allure report on GitLab pages including the tables with results, pipeline links and navigation between the different Allure reports.
• Github Action - report generation + historic reports

Troubleshooting

Answers to common questions/issues:

• I open allure report and I see just empty results with NaN counters

  You should not open allure report directly as a static html page. It uses local resources, thus is banned by modern browsers and requires web server to be opened properly. To resolve it you can disable CORS (not recommended), use live server extension for vs code, or just use allure serve command (recommended). To serve generated report an s3 bucket with hosting option could be used or any other web hosting.

• My other plugins do not work / allure-results is not generated

  It is likely other plugins (as cucumber-preprocessor) may also listen to events (especially after:spec) in Cypress that this plugin uses. Unfortunately, only one listener is available and other are just overwritten, that's why you can pass env variable allureReuseAfterSpec: true to not create new listeners from this plugin, but reuse existing. You can also try out cypress-on-fix plugin to register multiple listeners from your plugins, then this env var is not required.

Debugging

• for in-browser information (cypress events, mocha events, allure events, data collecting) execute localStorage.debug = 'allure-plugin*' in DevTools console
• for writer task information (writing results to disk, handling attachments, plugin events) add DEBUG=allure-plugin* before cypress run\open command

API

There are three options of using allure api inside tests:

1. Using interface from Cypress.Allure.reporter.getInterface() - synchronous

const allure = Cypress.Allure.reporter.getInterface();
allure.feature('This is our feature');
allure.epic('This is epic');
allure.issue('google', 'https://google.com');

2. Using Cypress custom commands, always starting from cy.allure() - chainer

cy.allure()
    .feature('This is feature')
    .epic('This is epic')
    .issue('google', 'https://google.com')
    .parameter('name', 'value')
    .tag('this is nice tag', 'as well as this');

3. Using Cypress-cucumber-preprocessor with cucumber tags:

@testID("id_of_test_for_testops")
@parentSuite("someParentSuite")
@suite("someSuite")
@subSuite("someSubSuite")
@epic("thisisepic")
@feature("nice")
@story("cool")
@severity("critical")
@owner("IAMOwner")
@issue("jira","JIRA-1234")
@tms("tms","TC-1234")
@link("other","url")
@someOtherTagsWillBeAddedAlso
Scenario: Here is scenario
...

Allure API available:

• testID(id: string)
• epic(epic: string)
• feature(feature: string)
• story(story: string)
• parentSuite(name: string)
• suite(name: string)
• subSuite(name:string)
• label(name: LabelName, value: string)
• parameter(name: string, value: string)
• testParameter(name: string, value: string)
• testName(name: string)
• link(url: string, name?: string, type?: LinkType)
• issue(name: string, url: string)
• tms(name: string, url: string)
• description(markdown: string)
• descriptionHtml(html: string)
• owner(owner: string)
• severity(severity: Severity)
• tag(tags: ...string)
• attachment(name: string, content: Buffer | string, type: ContentType)
• testAttachment(name: string, content: Buffer | string, type: ContentType)
• fileAttachment(name: string, path: string, type: ContentType)
• startStep(name: string)
• endStep()
• step(name: string, isParent: boolean)
• logStep(name: string)
• logCommandSteps(state: boolean)

Cypress commands as steps

Commands are producing allure steps automatically based on cypress events and are trying to represent how code and custom commands are executed with nested structure.
Moreover, steps functionality could be expanded with:

• cy.allure().step('name') - will create step "name" for current test. This step will be finished when next such step is created or test is finished.
• cy.allure().step('name', false) OR cy.allure().logStep('name') - will create step "name" for current parent step/hook/test. Will be finished when next step is created or test finished.
• cy.allure().startStep('name') - will create step "name" for current cypress command step / current step / current parent step / current hook or test. Is automatically finished on fail event or test end, but I would recommend to explicitly mention cy.allure().endStep() which will finish last created step.

To disable tracking of specific cypress commands to be not logged as steps in allure you can set env variable allureAvoidLoggingCommands which should contain an array of command names to be ignored, for example:

allureAvoidLoggingCommands: ["intercept", "myCustomCommand"]

To disable tracking of all cypress commands for specific code block you can use logCommandSteps api method:

// disable tracking cypress commands:
cy.allure().logCommandSteps(false);
cy.login(username, password);
// enable tracking cypress commands back again:
cy.allure().logCommandSteps();

Screenshots and Videos

Screenshots are attached automatically, for other type of content feel free to use testAttachment (for current test), attachment (for current executable), fileAttachment (for existing file).

Videos are attached for failed tests only from path specified in cypress config videosFolder and in case you have not passed video=false to Cypress configuration. In case you want to attach videos for passed tests please use allureAddVideoOnPass=true env variable.

It is done with the help of After Spec API. It will be used for:

• run mode with v6.7.0 and above
• run mode with v6.2.0 and above (but below v6.7.0) with experimentalRunEvents enabled
• interactive (open) mode for v7.1.0 with experimentalInteractiveRunEvents enabled When one of this conditions is satisfied - after:spec event will be used for attachments. It will reliably copy all screenshots available for each test and video (if available) to your allure-results folder and attach to each of your tests, so you don't need to somehow upload your videos and configure paths, etc.

In lower versions some other heuristics would be used, but they are not as reliable as after:spec.

Test name duplicates

By default Allure calculates hash from test title to identify test and show its' proper previous results.
This may lead to tests having the same name being counted by allure as retries of the same test. There are several ways to avoid this situation:

• the best way to avoid it is basically using unique test names

• update specific test name

      cy.allure().testName('new_test_name')

• specify your own function for all tests to not only take test.title, but also concatenate it with some other information in cypress/support/index or cypress/support/e2e.js file, for example:

  • use relative spec file path like "cypress/e2e/results2/test.cy.js" and test title:

    Cypress.Allure.reporter.getInterface().defineHistoryId((title) => {
        return `${Cypress.spec.relative}${title}`;
    });

  • use browser name and test title:

    Cypress.Allure.reporter.getInterface().defineHistoryId((title) => {
        return `${Cypress.browser.name}${title}`;
    });

  The rule is that this function should return any string (folder name, project name, platform, browser name, Cypress.spec content, etc.), and if those strings will be different - their test historyId hashes will be different - tests will be recognized as different by allure.

Suite structuring

Allure support 3 levels of suite structure:

• Suite tab: parentSuite -> suite -> subSuite -> tests
• Behaviors tab: epic -> feature -> story -> tests

They are defined automatically by structure passed from cypress mocha test object with titles of each parent. So an array of names of describe blocks is just transformed into: [parentSuite, suite, "subsuite1 -> subsuite2 -> ..."]

However, since v2.29.0 you can modify the strategy of defining names for structuring the tests by overwriting the function in support/index file using Cypress.Allure.reporter.getInterface().defineSuiteLabels which will accept your function:

// remove all describe block names and leave just last one:
Cypress.Allure.reporter
    .getInterface()
    .defineSuiteLabels((titlePath, fileInfo) => {
        return [titlePath.pop()];
    });

This function will have 2 arguments. titlePath is that array of describe names, and fileInfo is a parsed representation of a filepath for cases when you want to include folder or filename into some names, or just wrap suites in folders, or implement any of your ideas how to structure tests in reports.

// supplement parentSuite name with folder name
Cypress.Allure.reporter
    .getInterface()
    .defineSuiteLabels((titlePath, fileInfo) => {
        const [parentSuite, suite, ...subSuites] = titlePath;
        return [`${fileInfo.folder} | ${parentSuite}`, suite, ...subSuites];
    });

// make folder name a parentSuite:
Cypress.Allure.reporter
    .getInterface()
    .defineSuiteLabels((titlePath, fileInfo) => {
        return [fileInfo.folder, ...titlePath];
    });

// remove any other describe blocks and just show last one:
Cypress.Allure.reporter
    .getInterface()
    .defineSuiteLabels((titlePath, fileInfo) => {
        return [titlePath.pop()];
    });

// remove describe names and just place tests in folder -> filename structure:
Cypress.Allure.reporter
    .getInterface()
    .defineSuiteLabels((titlePath, fileInfo) => {
        return [fileInfo.folder, fileInfo.name];
    });

Gherkin and links

It is posible to pass tms link or issue link with tags tms("ABC-111") and issue("ABC-222"). However, that will not work well with Scenario Outlines which may have different examples being linked to different tasks or test cases in tms. So, plugin will also parse your scenario outlines with examples and in case header in table will be tms or issue - it will add it as link to report.

    Scenario Outline: Some scenario
        Given User want to link test <number> to tms
        When User creates examples table with tms and issue headers
        Then Links will be added to allure
        Examples:
            | tms    | issue   | number |
            | TEST-1 | JIRA-11 | 1      |
            | TEST-2 | JIRA-22 | 2      |

VS Code Helper plugin for cypress + cucumber

In case you are using VS Code and Cypress Helper (latest) extension, it has configuration for allure cucumber tags autocompletion in feature files:

"cypressHelper.cucumberTagsAutocomplete": {
        "enable": true,
        "allurePlugin": true,
        "tags": ["focus", "someOtherTag"]
    }

Credits

Thanks to Serhii Korol who made Allure-mocha reporter. Integration with Cypress internal mocha runner was based on that solution.

License

Copyright 2020-2023 Oleksandr Shevtsov [email protected].
This project is licensed under the Apache 2.0 License.